1. Field of the Invention
Embodiments of the specification generally relate to wireless communications and more specifically to a negative amplitude polar transmitter.
2. Description of the Related Art
Wireless communications systems generally use radio frequency (RF) signals to transmit data from a transmitter to one or more receivers. Wireless communication systems are frequently used to implement wireless local area networks (LANs) in which data is transmitted and received between computers, servers, Ethernet switches, hubs, and the like. A wireless LAN may, for example, enable web page data to be transferred between a server and a computer.
Wireless communication systems often transmit data through transmitters that use traditional heterodyne architectures. These architectures typically involve the creation of Cartesian (I and Q) waveforms and then up-mixing the waveforms to a desired frequency. Heterodyne architectures, unfortunately, may require many processing units to handle the Cartesian waveforms, such as a plurality of low-pass filters, baseband amplifiers, mixers and a linear RF amplifier.
Polar transmission architectures may reduce the size and power consumption of a transmitter by, among other things, removing one or more up-mixing stages from the transmitter. Polar transmitters are typically configured to transmit data based upon amplitude and phase waveforms rather than Cartesian I and Q waveforms. The amplitude typically ranges from zero to one (i.e. amplitude values may be normalized) and the phase values typically range from zero to +/−π. The phase values are cyclic with a period of 2π. In one embodiment, the derivative of phase data with respect to time is frequency. Thus, a change of a phase value of π represents the maximum frequency deviation that the polar transmitter may produce.
FIG. 1 is a conceptual diagram of a standard polar transmitter 100, which includes a polar amplifier 101 and a frequency synthesizer 102. The phase waveform may be determined by frequency synthesizer 102 and the amplitude waveform may be determined by amplitude data. Inputs to frequency synthesizer 102 may be modulating data and carrier frequency data. In one embodiment, the carrier frequency may be between 2402 and 2480 MHz and the modulating data may be +/−4 MHz, thereby enabling polar transmitter 100 to transmit frequencies commonly used by Bluetooth™ equipment (which may be specified by Bluetooth™ Special Interest Group). The output of frequency synthesizer 102 is a modulated frequency waveform. Both the modulated frequency waveform and the amplitude data are provided to the inputs of polar amplifier 101, which modulates both the modulated frequency waveform and the amplitude data and constructs an RF signal for transmission. In one embodiment, polar transmitter 100 may be implemented, in part, as a discrete-time system.
One drawback of polar transmitter 100 is that relatively large changes in frequency may be difficult to accurately produce within frequency synthesizer 102. In some cases, frequency synthesizer 102 may not be able to produce the required frequency. As a result, polar transmitter 100 may transmit a distorted or erroneous signal, which may reduce the overall performance of the transmitter. In the case of polar transmitters, a relatively large frequency change is one in which the difference between phase values of two polar transmission points is relatively close to π. Phase values are described in greater detail below in FIGS. 2A and 2B.
FIGS. 2A and 2B are graphs illustrating two polar transmission points for a polar transmitter. FIG. 2A illustrates a first transmission point y, which may be described in polar form by amplitude A and phase value θ. FIG. 2B illustrates a second transmission point ŷ. In this example, point ŷ has a different amplitude B and a different phase value compared to point y. As shown in FIG. 2B, the phase value of point ŷ is offset by π from phase value of point y. Therefore, point ŷ represents the maximal frequency deviation possible from point y.
Points y and ŷ may represent two transmission points of a discrete-time polar transmitter. As is well-known, many discrete-time systems may be configured to process data in an oversampled manner. An oversampled discrete-time system may use a relatively greater sampling frequency (and therefore relatively more samples) to process data than the fewest number of samples required by the Nyquist theorem (e.g. an eight times oversampled system may use eight times the Nyquist sample rate to process data). If an oversampled polar transmitter is used to transmit the points y and ŷ, then additional points between y and ŷ may be transmitted as well. These additional points are shown as squares in FIG. 2B.
FIGS. 3A, 3B, and 3C are graphs illustrating amplitude, phase and frequency waveforms, respectively, that may occur when moving from a first transmission point to a second polar transmission point using the discrete-time polar transmitter of FIG. 1. In this example, the first polar transmission point is point y in FIG. 2A and the second polar transmission point is point ŷ in FIG. 2B.
The amplitude waveform of FIG. 3A shows the amplitude of point y (A) and the amplitude of point ŷ (B). FIG. 3A also shows additional amplitude points that may be due to oversampling as the amplitude decreases from A, goes to zero, and increases to B. As shown in the phase waveform of FIG. 3B, the difference between the phase value at point y and the phase value at point ŷ (i.e. the phase transition) is π.
Note that the phase transition is also reflected as a frequency change (see FIG. 3C) because, in one embodiment, the derivative of the phase data is frequency data. Further, because the phase transition is a step function, the frequency change is an impulse function. In a discrete time system, the width of the impulse may be 1/(sampling period T) as shown in the FIG. 3C.
As indicated above, one drawback of a polar transmitter is that it may be relatively difficult to process relatively large frequency changes. In particular, a frequency synthesizer may not be able to track relatively a large frequency change, such as that shown in FIG. 3C. Furthermore, as the sampling rate increases, the height of the impulse in FIG. 3C increases, thereby indicating relatively greater frequency change requirements in relatively shorter times. Thus, relatively higher sampling rates require a frequency synthesizer to be able to track relatively greater changes in frequency.
Therefore, what is needed is a polar transmitter architecture that can support relatively large changes in phase and frequency.